Power grid load scheduling method and device, electronic equipment and storage medium

Abstract

The invention discloses a power grid load scheduling method and device, electronic equipment and a storage medium. The method comprises the steps that power grid operation data are acquired, and the power grid operation data are used for representing data related to power grid loads; constructing load characteristics of the power grid based on the power grid operation data; inputting the load characteristics into a target prediction model, and predicting load data of the power grid in a future time period by using the target prediction model to obtain a prediction result of the power grid load; and scheduling the power grid operation based on the prediction result. The technical problems of slow response and poor accuracy of power grid load scheduling in related technologies are solved.

Description

Technical Field

[0001] The present invention relates to the field of power system and power grid dispatching, and in particular to a method, device, electronic equipment and storage medium for dispatching power grid load. Background Art

[0002] During the operation of the power system, it is necessary to maintain a balance between power supply and demand. If the load on the power grid exceeds the power supply capacity, it may cause the power grid frequency to drop and the voltage to be unstable. On the contrary, if the power supply is excessive, it will not only waste resources but also increase the cost of power generation. Therefore, it is necessary to carry out power grid load dispatching to ensure real-time matching of power supply and demand and maintain the stable operation of the power grid.

[0003] At present, power grid load dispatching mainly relies on dispatchers' analysis based on statistics and mathematical models. Due to the limitations of dispatchers' knowledge and abilities, dispatchers' decisions have certain limitations, resulting in slow dispatching response and poor accuracy of power grid load.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a method, device, electronic device and storage medium for scheduling power grid load, so as to at least solve the technical problems of slow response and poor accuracy of power grid load scheduling in related technologies.

[0006] According to one aspect of an embodiment of the present invention, there is provided a method for scheduling power grid load, comprising: acquiring power grid operation data, wherein the power grid operation data is used to characterize data related to the power grid load; constructing load characteristics of the power grid based on the power grid operation data; inputting the load characteristics into a target prediction model, using the target prediction model to predict the load data of the power grid in a future time period, and obtaining a prediction result of the power grid load; and scheduling the power grid operation based on the prediction result.

[0007] Optionally, obtaining the grid operation data includes: collecting initial grid operation data; and preprocessing the initial grid operation data to obtain the grid operation data.

[0008] Optionally, the initial operation data is preprocessed, including: determining that there is missing blank data in the initial operation data, and filling the blank data; determining that there is abnormal data in the initial operation data, and correcting the abnormal data.

[0009] Optionally, based on the power grid operation data, load characteristics are constructed, including: extracting a load characteristic subset with a correlation greater than a preset threshold from the power grid operation data through correlation analysis; performing time series analysis on the power grid operation data to obtain the time characteristics of the power grid load; identifying the peak value in the power grid operation data through cross-enrichment following to obtain the peak characteristics of the power grid load; constructing initial load characteristics based on the load characteristic subset, time characteristics and peak characteristics; and evaluating the initial load characteristics through analog source-flow interaction to obtain the load characteristics.

[0010] Optionally, a time series analysis is performed on the power grid operation data to determine the time characteristics, including: randomly selecting multiple time points from the power grid operation data and considering the time points as cluster centers; based on the cluster centers, clustering multiple time points in a preset time period by a clustering algorithm to obtain clustering results, wherein the clustering results are used to characterize multiple types of power grid load change patterns; based on the clustering results, determining the time characteristics.

[0011] Optionally, based on the clustering results, determining the time characteristics includes: based on the clustering results, determining an initial time interval from the power grid operation data; and denoising and smoothing the initial time interval to obtain the time characteristics.

[0012] Optionally, the above method also includes: dividing a verification set from the power grid operation data according to a preset ratio; based on the verification set, adjusting the initial prediction model to obtain a target prediction model.

[0013] Optionally, the initial prediction model is adjusted, including: adjusting the parameters of the initial prediction model by integrating counter-proof and error analysis based on the validation set; adjusting the structure of the initial prediction model by node comparison based on the validation set.

[0014] Optionally, the operation of the power grid is dispatched based on the prediction results, including: generating multiple dispatch strategies for the power grid based on the prediction results; determining a target strategy from multiple dispatch strategies through preset evaluation rules; and dispatching the operation of the power grid based on the target strategy.

[0015] According to another aspect of an embodiment of the present invention, there is also provided a scheduling device for power grid load, comprising: an acquisition module for acquiring power grid operation data, wherein the power grid operation data is used to characterize data related to the power grid load; a construction module for constructing the load characteristics of the power grid based on the power grid operation data; a prediction module for inputting the load characteristics into a target prediction model, and using the target prediction model to predict the load data of the power grid in a future time period to obtain a prediction result of the power grid load; and a scheduling module for scheduling the power grid operation based on the prediction result.

[0016] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present invention is executed when the program is running.

[0017] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0018] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0019] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0020] According to another aspect of the embodiments of the present invention, a computer program is further provided. When the computer program is executed by a processor, the methods in the embodiments of the present invention are implemented.

[0021] In an embodiment of the present invention, power grid operation data is obtained, wherein the power grid operation data is used to characterize data related to the power grid load; based on the power grid operation data, the load characteristics of the power grid are constructed; the load characteristics are input into the target prediction model, and the target prediction model is used to predict the load data of the power grid in the future time period to obtain the prediction result of the power grid load; and the power grid operation is scheduled based on the prediction result. It is easy to notice that the method of scheduling the power grid operation based on the power grid load prediction result is to extract the load characteristics from the power grid operation data so that the load characteristics can accurately reflect the power grid load changes, and further predict the prediction results based on the load characteristics. The prediction results can accurately reflect the load data in the future time period, so that the power grid operation can be effectively scheduled using the prediction results, and the purpose of effectively scheduling the power grid load is achieved, thereby achieving the technical effect of improving the efficiency of power grid load scheduling, and then solving the technical problems of slow response and poor accuracy of power grid load scheduling in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0023] Figure 1is a flow chart of a method for dispatching power grid load according to an embodiment of the present invention;

[0024] Figure 2 is a schematic diagram of a dispatching system for power grid loads according to an embodiment of the present application;

[0025] Figure 3 is a schematic diagram of a data source collection module in a power grid load dispatching system according to an embodiment of the present application;

[0026] Figure 4 is a schematic diagram of a feature capture and selection module in a power grid load dispatching system according to an embodiment of the present application;

[0027] Figure 5 is a schematic diagram of an evaluation and calibration module in a dispatching system for power grid loads according to an embodiment of the present application;

[0028] Figure 6 is a flow chart of an optional method for predicting power grid load according to an embodiment of the present application;

[0029] Figure 7 is a schematic diagram of a dispatching device for power grid load according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0032] According to an embodiment of the present invention, an embodiment of a method for dispatching power grid loads is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0033] Figure 1 is a flow chart of a method for dispatching power grid load according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0034] Step S102, obtaining power grid operation data.

[0035] Among them, the power grid operation data is used to characterize the data related to the power grid load.

[0036] The above-mentioned grid load can refer to the total electric power consumed by the power equipment in the grid. The grid load can be a conventional load, a peak load or a variable load. Among them, the conventional load is generally periodic. Excessive peak load may cause grid overload, affecting the stability and security of power supply. Therefore, it is necessary to analyze the grid load for grid load scheduling, which is also helpful for grid planning and construction to meet future electricity demand. The grid load will be affected by power equipment or weather.

[0037] The above-mentioned power grid operation data can be used to characterize data related to power grid load. Power grid operation data may include but are not limited to: load data, environmental data, power generation data, etc., and the power grid operation data is not limited here and can be determined as needed. Specifically, load data may include but are not limited to: electric power, current, etc., and the load data is not limited here and can be determined as needed. Environmental data may include but are not limited to: temperature, humidity, light, etc., and the environmental data is not limited here and can be determined as needed. Power generation data may include but are not limited to: power generation, etc., and the power generation data is not limited here and can be determined as needed. By analyzing the power grid operation data, the characteristics that affect the power grid load can be obtained, and then the power grid load can be predicted based on the power grid load characteristics, and further based on the future power grid load, a dispatching strategy can be formulated to save power resources. Power grid operation data is the basis of power grid load characteristics and an important reference for power grid load dispatching.

[0038] In an optional embodiment, the grid operation data can be collected in real time. Specifically, the grid load data can be monitored and recorded by smart meters, the environmental data can be monitored by sensors, and the power generation data can be obtained by the automatic control system in the grid system. In another optional embodiment, the grid operation data can also be obtained from a cloud storage platform. Specifically, the grid operation data can be stored in the cloud storage platform in advance. When performing scheduling analysis, the grid operation data is retrieved from the cloud storage platform. The grid operation data can be based on multi-source data, thereby ensuring the effectiveness of charge feature extraction, so as to facilitate the accuracy of subsequent grid load prediction and scheduling.

[0039] Step S104: constructing the load characteristics of the power grid based on the power grid operation data.

[0040] The above-mentioned load characteristics can be variables that affect the change of grid load. The correlation with grid load can be analyzed from the grid operation data, so as to screen out load characteristics with correlation greater than the preset threshold. Through load characteristics, it is easy to understand the dynamic changes of grid load and predict the power demand in the future time period, so as to improve the utilization rate of the grid. Through load characteristics, it is helpful to further learn the pattern and law of grid load change through the prediction model. The selection and construction of load characteristics help to improve the accuracy, stability and computational efficiency of the prediction model. Accurate load characteristics can not only capture the law of grid load change, but also enhance the robustness and adaptability of the prediction model. And in the actual operation of the power grid, the load characteristics will be continuously updated and adjusted with the changes in the scale of the power grid and user behavior.

[0041] In an optional embodiment, the load characteristics of the power grid can be constructed by a statistical analysis method. Specifically, first, multiple candidate features can be extracted from the power grid operation data, and then the correlation between the multiple candidate features and the power grid load is quantitatively evaluated. Optionally, the influence of multiple candidate features on the power grid load is evaluated by statistical indicators such as the Pearson correlation coefficient and the Gini index. Then, by setting a preset threshold, features whose correlation with the power grid load is less than the preset threshold are excluded, and a preliminary feature set is constructed. Secondly, experts can screen out the load characteristics of the power grid from the preliminary feature set.

[0042] In another optional embodiment, the load characteristics can also be selected by a machine learning algorithm. The machine learning algorithm can be one of the following: random forest, decision tree model, and the machine learning algorithm is not limited here and can be determined as needed. Through the random forest algorithm, the load characteristics can be predicted and the importance score of the load characteristics can be determined. Furthermore, a recursive feature elimination method can be used to gradually remove features whose contribution is less than a preset contribution, so as to achieve the simplification of the load characteristics.

[0043] In another optional embodiment, a load feature subset with a correlation greater than a preset threshold can be extracted from the power grid operation data through correlation analysis. Then, the time characteristics of the power grid load are obtained through time series analysis, and the peak characteristics of the power grid load are obtained by cross-enrichment following to identify the peak in the power grid operation data. Further, the obtained load feature subset, time characteristics and peak characteristics are evaluated to obtain the final load characteristics.

[0044] Step S106, inputting the load characteristics into the target prediction model, using the target prediction model to predict the load data of the power grid in the future time period, and obtaining the prediction result of the power grid load.

[0045] The above-mentioned target prediction model can be a model for predicting the grid load required for the future time period of the grid. The prediction model can be pre-trained. The target prediction model can be a statistical model, such as linear regression, time series analysis, etc., which makes predictions based on the statistical laws of historical data. The target prediction model can also be a machine learning model, such as a support vector machine, a random forest, a deep learning model, etc. The target prediction model is not limited here and can be determined as needed. The target prediction model can predict the load data of the grid in the future time period based on the load characteristics.

[0046] The above-mentioned prediction may refer to the process of predicting the load data of the power grid in the future time period, so as to facilitate the planning of power resources for early allocation and scheduling. By accurately predicting the future load data of the power grid, it is easy to understand the future power demand, so as to reasonably arrange the power generation plan, improve the power generation scheduling, reduce the power generation cost and improve the power supply efficiency. By accurately predicting the future load data of the power grid, it is also possible to predict potential peak power consumption in advance, prepare for emergencies, adjust the power supply strategy in time, and reduce the risk of power outages.

[0047] The above-mentioned prediction results can be used to characterize the required grid load conditions in future time periods. The prediction results may include, but are not limited to: load value, load change trend, etc. The prediction results are not limited here and can be determined as needed. The prediction results can be obtained by the target prediction model based on the load characteristics prediction. Through the prediction results, the demand for the grid in the future time period can be understood, and the supply and demand balance of the power system can be planned, so that when faced with fluctuations in load forecasts, it can respond quickly to adjust the grid load and ensure the reliability of power supply.

[0048] In an optional embodiment, an autoregressive integral moving average model can be used to predict the load on the power grid. Specifically, the load characteristics can be standardized and then input into the autoregressive integral moving average model. The autoregressive integral moving average model captures the time series dependency of the power grid load through autoregressive, differential, and moving average components, predicts the load data of the power grid in the future time period, and obtains the prediction result of the power grid load. The prediction result can help the dispatching center adjust the power generation plan in advance to ensure the balance between supply and demand, especially to take measures in advance to adjust the load changes faster, such as the load surge in the early morning or the load peak in the evening.

[0049] In another optional embodiment, a random forest algorithm can be used to predict the power grid load. The target prediction model is obtained by training the random forest algorithm, which is composed of multiple decision trees, so that the prediction result is a weighted average of the prediction values ​​of multiple decision trees, which can reduce the risk of overfitting and improve the robustness of the prediction.

[0050] Step S108: scheduling the operation of the power grid based on the prediction results.

[0051] The above-mentioned dispatching refers to the process of distributing the load of the power grid. Dispatching can be real-time dispatching or dispatching in the future time period. By dispatching the operation of the power grid, it is possible to better allocate power resources and improve the reliability of the power system to cope with the power demand in the future time period.

[0052] In an optional embodiment, the prediction results can be analyzed to determine the required grid load in the future time period. Then, a scheduling strategy is generated for the required grid load so that power resources are allocated based on the scheduling strategy. Optionally, the scheduling strategy may include but is not limited to: adjusting the power output of power equipment, allocating the power flow of transmission lines, and setting the charging and discharging timing of energy storage equipment. The scheduling strategy is not limited here and can be determined as needed.

[0053] In another optional embodiment, multiple dispatching strategies for power grid operation can be generated based on the prediction results. Optionally, multiple dispatching strategies can be generated through genetic algorithms, particle swarm optimization or linear programming. Then, through preset evaluation rules, the target strategy is determined from multiple dispatching strategies to dispatch the power grid operation according to the current prediction results. Through the prediction results, the changing trend of the power grid load can be predicted, providing a decision-making basis for power dispatching, effectively avoiding power grid overload, and ensuring the stable operation of the power system.

[0054] In an embodiment of the present invention, power grid operation data is obtained, wherein the power grid operation data is used to characterize data related to the power grid load; based on the power grid operation data, the load characteristics of the power grid are constructed; the load characteristics are input into the target prediction model, and the target prediction model is used to predict the load data of the power grid in the future time period to obtain the prediction result of the power grid load; and the power grid operation is scheduled based on the prediction result. It is easy to notice that the method of scheduling the power grid operation based on the power grid load prediction result is to extract the load characteristics from the power grid operation data so that the load characteristics can accurately reflect the power grid load changes, and further predict the prediction results based on the load characteristics. The prediction results can accurately reflect the load data in the future time period, so that the power grid operation can be effectively scheduled using the prediction results, and the purpose of effectively scheduling the power grid load is achieved, thereby achieving the technical effect of improving the efficiency of power grid load scheduling, and then solving the technical problems of slow response and poor accuracy of power grid load scheduling in related technologies.

[0055] Optionally, obtaining the grid operation data includes: collecting initial grid operation data; and preprocessing the initial grid operation data to obtain the grid operation data.

[0056] The above-mentioned initial operation data can be used to characterize data related to the power grid load. The initial operation data may include but is not limited to: load data, environmental data, power generation data, etc., and the initial operation data is not limited here and can be determined as needed. Specifically, the load data may include but is not limited to: electric power, current, etc., and the load data is not limited here and can be determined as needed. Environmental data may include but is not limited to: temperature, humidity, light, etc., and the environmental data is not limited here and can be determined as needed. Power generation data may include but is not limited to: power generation, etc., and the power generation data is not limited here and can be determined as needed. The initial operation data can provide data resources for the subsequent capture of the load characteristics of the power grid, so that the load data of the power grid in the future time period can be predicted.

[0057] In an optional embodiment, the initial operation data of the power grid can be monitored and recorded by a monitoring device. The monitoring device may include, but is not limited to, electric meters, sensors, etc. The monitoring device is not limited here and can be determined as needed. Due to measurement errors or network interference, the initial operation data may be abnormal, so it is necessary to preprocess the initial operation data to obtain the power grid operation data. The preprocessing process not only improves the reliability of the power grid operation data, but also facilitates subsequent predictions through the target prediction model, thereby improving the efficiency and accuracy of power system scheduling. Optionally, the missing values ​​in the initial operation data can be filled based on the interpolation of the time series, the initial operation data can be standardized, or the abnormal values ​​in the initial operation data can be identified and corrected through a machine learning model.

[0058] Optionally, the initial operation data is preprocessed, including: determining that there is missing blank data in the initial operation data, and filling the blank data; determining that there is abnormal data in the initial operation data, and correcting the abnormal data.

[0059] In an optional embodiment, the initial operation data may be preprocessed by blank data processing and abnormal data processing. Specifically, blank data processing may first determine the blank data, and then fill in the blank data. Optionally, the blank data may be filled based on time series or historical data. The filling method is not limited here and can be determined as needed. Abnormal data processing may first determine the abnormal data, and then correct the abnormal data. Optionally, the correction of abnormal data may be manually processed by staff, or may be corrected by statistical methods or machine learning methods. The correction method is not limited here and can be determined as needed. Preprocessing can effectively solve the problems of missing and abnormal data, and improve the reliability and prediction accuracy of the target prediction model.

[0060] Optionally, based on the power grid operation data, load characteristics are constructed, including: extracting a load characteristic subset with a correlation greater than a preset threshold from the power grid operation data through correlation analysis; performing time series analysis on the power grid operation data to obtain the time characteristics of the power grid load; identifying the peak value in the power grid operation data through cross-enrichment following to obtain the peak characteristics of the power grid load; constructing initial load characteristics based on the load characteristic subset, time characteristics and peak characteristics; and evaluating the initial load characteristics through analog source-flow interaction to obtain the load characteristics.

[0061] In an optional embodiment, a subset of load characteristics is extracted by correlation analysis. Optionally, the correlation can be analyzed statistically, for example, by statistical indicators such as the Pearson correlation coefficient and the Gini index. The correlation can also be analyzed by a machine learning model. In this way, a subset of load characteristics with a correlation greater than a preset threshold can be extracted from the power grid operation data to further analyze the load characteristics. Since periodic or peak power consumption is usually present in the operation of the power grid, it is necessary to conduct an in-depth analysis of the time characteristics and peak characteristics to obtain accurate load characteristics. Specifically, the power grid operation data can be subjected to a time series analysis to obtain the time characteristics of the power grid load. The peak value in the power grid operation data can be identified by cross-enrichment follow-up to obtain the peak characteristics of the power grid load. Then, the constructed initial load characteristics are evaluated, and the characteristics with higher importance are selected to obtain the load characteristics. Constructing load characteristics can comprehensively capture the dynamic changes and peak characteristics of the power grid load, and provide a data basis for predicting the power grid load in the future time period.

[0062] Among them, cross-enrichment following can enhance the recognition ability of load peaks through multi-source data, and can extract peak features from multi-source power grid operation data, so that the target prediction model can provide more robust and reasonable prediction results when facing new load characteristics or rare events.

[0063] Among them, analog source flow interaction can be achieved by analyzing the dynamic interaction between historical and current events similar to the current situation and data flow, and based on the development trajectory of the analog source and its comparative analysis with the current situation. Therefore, the initial load characteristics can be evaluated through analog source flow interaction, and the characteristics with higher importance can be selected from the initial load characteristics to obtain the load characteristics.

[0064] Optionally, a time series analysis is performed on the power grid operation data to determine the time characteristics, including: randomly selecting multiple time points from the power grid operation data and considering the time points as cluster centers; based on the cluster centers, clustering multiple time points in a preset time period by a clustering algorithm to obtain clustering results, wherein the clustering results are used to characterize multiple types of power grid load change patterns; based on the clustering results, determining the time characteristics.

[0065] In an optional embodiment, the time characteristics may be determined by clustering. Through time series analysis, the periodicity and seasonality of the grid load may be found, providing the time dimension characteristics for the target prediction model.

[0066] Firstly, multiple time points are randomly selected from the power grid operation data, and these multiple time points are respectively used as multiple cluster centers for clustering.

[0067] Then, in order to analyze the grid load characteristics of the preset time period, multiple time points in the preset time period are clustered and analyzed to obtain clustering results, which can be used to characterize multiple types of grid load change patterns. Specifically, the similarity between each time point in the preset time period and these multiple cluster centers can be calculated respectively. For example, the similarity can be calculated by Euclidean distance, cosine similarity or correlation coefficient, and the time points with the closest similarity are clustered to the corresponding cluster centers until all time points in the preset time period are clustered to the corresponding cluster centers, and finally the clustering results are obtained. Different cluster centers can reflect different patterns and trends of grid load.

[0068] Optionally, based on the clustering results, determining the time characteristics includes: based on the clustering results, determining an initial time interval from the power grid operation data; and denoising and smoothing the initial time interval to obtain the time characteristics.

[0069] By clustering, dynamic time series analysis can be achieved, which helps to automatically identify non-fixed patterns or changing trends in power grid load, such as sudden upward or downward trends. Further denoising and smoothing of the initial time interval can focus on long-term trends or stable patterns of seasonal repetition in the time series, thereby eliminating noise, smoothing power grid load changes, and improving the stability and accuracy of the target prediction model.

[0070] Optionally, the method further includes: dividing a verification set from the power grid operation data according to a preset ratio; and adjusting the initial prediction model based on the verification set to obtain a target prediction model.

[0071] The above validation set can be used to adjust the initial prediction model and improve the prediction accuracy. To improve efficiency, a validation set is divided from the power grid operation data to adjust the model. The validation set can be used to evaluate the model performance and adjust the initial prediction model. Through the validation set, the initial prediction model can be adjusted to obtain the target prediction model to improve the applicability of the target prediction model.

[0072] Optionally, the initial prediction model is adjusted, including: adjusting the parameters of the initial prediction model by integrating counter-proof and error analysis based on the validation set; adjusting the structure of the initial prediction model by node comparison based on the validation set.

[0073] In an optional embodiment, the initial prediction model can be evaluated and adjusted based on the validation set, so that the parameters and structure of the initial prediction model can be optimized to improve the adaptability and prediction accuracy of the target prediction model.

[0074] The generalization ability of the initial prediction model can be evaluated by error analysis. Specifically, the mean square error can be obtained, that is, the error between the predicted value and the true value of the initial prediction model, and the parameters of the initial prediction model can be adjusted based on the error until the error is less than a preset error threshold.

[0075] Integrated disproof can be used to identify potential problems and deviations in the initial prediction model. Integrated disproof can search for evidence and data points that contradict the prediction results through the method of contradiction. Further, the strength of disproof can be used to measure the ratio of the predicted value of the initial prediction model to the true value, thereby evaluating the effectiveness of integrated disproof.

[0076] The structure of the initial prediction model can be tested through node comparison. Node comparison can help identify the risk points of overfitting or underfitting by comparing the contribution and consistency of different nodes and the generated features in the prediction process.

[0077] Optionally, the operation of the power grid is dispatched based on the prediction results, including: generating multiple dispatch strategies for the power grid based on the prediction results; determining a target strategy from multiple dispatch strategies through preset evaluation rules; and dispatching the operation of the power grid based on the target strategy.

[0078] The above-mentioned dispatching countermeasures can be used to dynamically adjust the operation strategy and resource allocation of the power grid. Since the prediction results can reflect the power demand and power generation potential in the future time period, multiple dispatching countermeasures for the power grid can be generated based on the prediction results. The dispatching countermeasures can be about power generation scheduling, or about transmission line scheduling, or about energy storage equipment scheduling. Specifically, for example, the dispatching countermeasures can be to adjust the power output of power equipment, allocate the power flow of transmission lines, or set the charging and discharging timing of energy storage equipment. The dispatching countermeasures are not limited here and can be determined as needed.

[0079] In an optional embodiment, multiple dispatching strategies for the power grid can be generated by genetic algorithm, particle swarm optimization or linear programming. Furthermore, according to the power distribution target or preset evaluation rules, the target strategy is determined from multiple dispatching strategies to adopt the optimal strategy to dispatch power resources. In this way, targeted dispatching strategies can be formulated based on the prediction results, effectively avoiding power grid overload and improving the operating efficiency of the power system.

[0080] Optionally, the preset evaluation rule can be determined by the following formula:

[0081] t=C1*(P 1t -P2 t ) 2 +C2*P2+C3*Et;

[0082] In the formula, t represents time, P 1t is the total power generation at time t, P 2t is the total electricity demand at the tth moment, P2 represents the cost of adjusting the power generation, Et is the energy state of the energy storage device, which can characterize the energy utilization efficiency, C1, C2, and C3 are the first weighting factor, the second weighting factor, and the third weighting factor, respectively.

[0083] Figure 2 is a schematic diagram of a dispatching system for power grid loads according to an embodiment of the present application, such as Figure 2 As shown, the scheduling system includes: a data source collection module, a feature capture and selection module, an evaluation and calibration module, and a dynamic control module.

[0084] Specifically, the data source collection module can be used to obtain power grid operation data. The feature capture and selection module can be used to extract load features from the power grid operation data. The evaluation and calibration module can be used to predict the power grid load in the future time period through the target load prediction model to obtain the prediction results, and can also be used to adjust the initial load prediction model to obtain the target load prediction model. The dynamic control module can be used to control the power grid operation based on the prediction results.

[0085] The data source collection module includes a data cleaning unit, an anomaly detection unit, an iterative feedback unit, and an incremental control unit. Figure 3 shown.

[0086] The data cleaning unit can be used to pre-process the collected original operation data of the power grid to obtain initial operation data, and can also send the initial operation data to the anomaly detection unit. The anomaly detection unit can be used to identify abnormal data in the initial operation data. The iterative feedback unit can be used to adjust the pre-processing process based on the data cleaning results and the anomaly detection results. The incremental control unit can be used to process the newly added operation data.

[0087] The feature capture and selection module includes a time series analysis unit, a peak load analysis unit and a prediction and classification feedback unit. Figure 4 shown.

[0088] The time series analysis unit can be used to analyze the time characteristics. Specifically, the dynamic and static sections are combined. The peak load analysis unit can be used to analyze the peak characteristics. Specifically, the cross-enrichment follow-up method is adopted. The prediction hierarchical feedback unit can be used to evaluate the characteristics. Specifically, the analog source flow interaction method is adopted.

[0089] The evaluation and calibration module includes an integrated counter-proof unit, a node comparison unit, a variable coefficient library and a prediction unit. Figure 5 shown.

[0090] The integrated counter-proof unit can be used to adjust the parameters of the initial load forecasting model. The node comparison unit can be used to adjust the structure of the load forecasting model. The variable coefficient library can be used to divide the verification set from the power grid operation data, and send the verification set to the integrated counter-proof unit and the node comparison unit, so as to adjust the parameters of the load forecasting model. The prediction unit can be used to predict the power grid load in the future time period using the target load prediction model to obtain the prediction result.

[0091] Figure 6 is a flow chart of an optional method for predicting power grid load according to an embodiment of the present application, such as Figure 6 As shown, the method includes:

[0092] First, collect multi-source data. Perform dynamic load feature selection. Specifically, select load features from multi-source data. Evaluate load features to ensure that load features accurately reflect grid load changes. Then, build and verify a prediction model. Specifically, build a prediction model based on load features and verify the model. And, dynamically adjust the output by comparing and disproving data sets in various ways.

[0093] According to an embodiment of the present invention, an embodiment of a scheduling device for power grid load is provided. It should be noted that the device can be used to execute the above-mentioned scheduling method for power grid load. The specific implementation scheme and application scenario in this embodiment are the same as those in the above-mentioned embodiment and will not be repeated here.

[0094] Figure 7 is a schematic diagram of a dispatching device for power grid load according to an embodiment of the present application, such as Figure 7 As shown, the device includes: an acquisition module 70, used to acquire power grid operation data, wherein the power grid operation data is used to characterize data related to the power grid load; a construction module 72, used to construct the load characteristics of the power grid based on the power grid operation data; a prediction module 74, used to input the load characteristics into a target prediction model, and use the target prediction model to predict the load data of the power grid in a future time period to obtain a prediction result of the power grid load; a scheduling module 76, used to schedule the power grid operation based on the prediction result.

[0095] Optionally, the acquisition module is also used to collect initial operation data of the power grid; pre-process the initial operation data of the power grid to obtain power grid operation data.

[0096] Optionally, the acquisition module is further used to determine whether there is missing blank data in the initial operation data and fill the blank data; determine whether there is abnormal abnormal data in the initial operation data and correct the abnormal data.

[0097] Optionally, the construction module is also used to extract a load feature subset with a correlation greater than a preset threshold from the power grid operation data through correlation analysis; perform time series analysis on the power grid operation data to obtain the time characteristics of the power grid load; identify the peak value in the power grid operation data through cross-enrichment following to obtain the peak characteristics of the power grid load; construct initial load characteristics based on the load feature subset, time characteristics and peak characteristics; evaluate the initial load characteristics through analog source-flow interaction to obtain the load characteristics.

[0098] Optionally, the construction module is also used to randomly select multiple time points from the power grid operation data and regard the time points as clustering centers; based on the clustering centers, multiple time points in a preset time period are clustered by a clustering algorithm to obtain clustering results, wherein the clustering results are used to characterize multiple types of power grid load change patterns; based on the clustering results, the time characteristics are determined.

[0099] Optionally, the construction module is also used to determine an initial time interval from the power grid operation data based on the clustering results; and perform denoising and smoothing on the initial time interval to obtain a time feature.

[0100] Optionally, the above device also includes a training module.

[0101] Optionally, a training module is used to divide a verification set from the power grid operation data according to a preset ratio; based on the verification set, the initial prediction model is adjusted to obtain a target prediction model.

[0102] Optionally, the training module is also used to adjust the parameters of the initial prediction model through integrated counter-evidence and error analysis based on the validation set; and adjust the structure of the initial prediction model through node comparison based on the validation set.

[0103] Optionally, the dispatching module is also used to generate multiple dispatching strategies for the power grid based on the prediction results; determine the target strategy from the multiple dispatching strategies through preset evaluation rules; and dispatch the power grid operation based on the target strategy.

[0104] An embodiment of the present application further provides an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the program executes the methods in various embodiments of the present invention when running.

[0105] An embodiment of the present application further provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0106] An embodiment of the present application further provides a computer program product, including a computer program, which implements the methods in various embodiments of the present invention when executed by a processor.

[0107] An embodiment of the present application further provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0108] The embodiments of the present application further provide a computer program, which implements the methods in the above-mentioned embodiments of the present invention when executed by a processor.

[0109] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0110] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0111] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0112] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0113] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0114] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for dispatching power grid load, characterized in that: include: Acquiring power grid operation data, wherein the power grid operation data is used to represent data related to the power grid load; Based on the power grid operation data, constructing the load characteristics of the power grid; The load characteristics are input into a target prediction model, and the target prediction model is used to predict the load data of the power grid in a future time period to obtain a prediction result of the power grid load; The operation of the power grid is scheduled based on the prediction result.

2. The method according to claim 1, characterized in that: Obtain power grid operation data, including: Collecting initial operation data of the power grid; The initial operation data of the power grid is preprocessed to obtain the operation data of the power grid.

3. The method according to claim 2, characterized in that The preprocessing of the initial operation data includes: Determining that there is missing blank data in the initial operation data, and filling the blank data; It is determined that abnormal data exists in the initial operation data, and the abnormal data is corrected.

4. The method according to claim 1, characterized in that: The constructing load characteristics based on the power grid operation data includes: Extracting a load feature subset having a correlation greater than a preset threshold from the power grid operation data through correlation analysis; Performing time series analysis on the power grid operation data to obtain the time characteristics of the power grid load; Identifying peak values ​​in the power grid operation data by cross-enrichment following to obtain peak characteristics of the power grid load; constructing an initial load feature based on the load feature subset, the time feature and the peak feature; The initial load characteristic is evaluated by analogy with source flow interaction to obtain the load characteristic.

5. The method according to claim 4, characterized in that The performing time series analysis on the power grid operation data to determine the time characteristics includes: Randomly selecting multiple time points from the power grid operation data, and considering the time points as cluster centers; Based on the cluster center, clustering multiple time points in a preset time period by a clustering algorithm to obtain a clustering result, wherein the clustering result is used to characterize multiple types of power grid load change patterns; Based on the clustering result, the time feature is determined.

6. The method according to claim 5, characterized in that The determining the time feature based on the clustering result includes: Based on the clustering result, determining an initial time interval from the power grid operation data; The initial time interval is subjected to denoising and smoothing processing to obtain the time feature.

7. The method according to claim 6, characterized in that Also includes: According to a preset ratio, a verification set is divided from the power grid operation data; Based on the validation set, the initial prediction model is adjusted to obtain a target prediction model.

8. The method according to claim 7, characterized in that The adjusting of the initial prediction model comprises: Based on the validation set, adjusting the parameters of the initial prediction model by integrating counter-evidence and error analysis; Based on the validation set, the structure of the initial prediction model is adjusted through node comparison.

9. The method according to claim 1, characterized in that: The scheduling of the power grid operation based on the prediction result includes: Based on the prediction results, generating a plurality of dispatching strategies for the power grid; Determining a target strategy from the plurality of scheduling strategies by using a preset evaluation rule; Based on the target strategy, the operation of the power grid is scheduled.

10. A dispatching device for power grid load, characterized in that: include: An acquisition module, used for acquiring power grid operation data, wherein the power grid operation data is used for representing data related to the power grid load; A construction module, used for constructing the load characteristics of the power grid based on the power grid operation data; A prediction module, used to input the load characteristics into a target prediction model, and use the target prediction model to predict the load data of the power grid in a future time period to obtain a prediction result of the power grid load; A scheduling module is used to schedule the operation of the power grid based on the prediction result.

11. An electronic device, characterized in that: include: A memory storing an executable program; A processor, configured to run the program, wherein the program executes the method according to any one of claims 1 to 9 when running.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is run, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 9.

13. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images